Post on 24-Apr-2020
1
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Full Vehicle Durability Prediction Using Co-simulation
Between Implicit & Explicit Finite Element Solvers
SIMULIA Great Lakes Regional User Meeting
Oct 12, 2011
Victor Oancea
Member of SIMULIA CTO Office
2
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Overview
Motivation
Co-simulation Vehicle Model
o Body
o Suspension
o Tires
Validation
Use cases
Suspension template
modeling in Abaqus/CAE K&C, Vibration, Durability
Use cases
Summary
3
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Motivation
Road Loads
Bench Test
Component Loads and Stresses
Fatigue Life Prediction
Vehicle Durability Workflow
4
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Motivation• Simulation can provide detailed insight into the vehicle behavior early in the design
cycle before physical prototypes are available
• Accurate and efficient simulation of vehicles on test track requires a broad range of functionality, like:
• High performance computing – Complex system level models, faster turnaround
Mechanisms Substructures(Linear) Plasticity and Failure Contact
5
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Co-Simulation
• Implicit vs. Explicit analysis for full vehicle durability
Implicit
• “Large” time increments;
each is relatively expensive
• Period of interest is long
relative to vibration
frequency
• Linear response
— Substructures
• Smooth nonlinear response
Explicit
• “Small” time increments;
each is inexpensive
• High-speed dynamics
• Discontinuous nonlinear
behavior
— Impact
— Material failure
6
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Co-simulation
• Identify the regions suited for Implicit and Explicit
solution techniques
• Explicit solution scheme for Tire-road
interaction— Rapid changes in contact state and impact
handled by the general contact algorithm in
Abaqus/Explicit
• Vehicle Body and Suspension solved
using Implicit solution scheme in
Abaqus/Standard— System of equations solved using HHT time
integration
Co-simulation
• The two parts are solved independently
• Individual solutions are coupled
together to ensure continuity of the
global solution across the interface
7
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Vehicle Model
• Substructures provide an efficient way
to model the linear response of the body
for long duration events
— Enhanced dynamic response with Fixed
interface modes, Free interface modes or Mixed
• High performance computing enables
full Body meshes with elastic-plastic
materials to be used in short duration
impact events where significant plastic
strains are expected
8
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Vehicle Model
Kinematics and Compliance simulation
• Kinematic joints and bushings in the suspension and steering subsystems modeled using 12 DOF connector elements
— Coupled behavior between various DOFs possible
— Bushing connectors calibrated using physical test data or detailed finite element models
— Friction, plasticity, damage and failure can be prescribed for the connector elements
• Suspension components modeled as rigid, substructure or non-linear deformable depending of the level of fidelity sought from the simulation
9
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Tire ModelRequirements: Accurate representation of spindle forces and moments
Impact with short wavelength obstacles
Ease of calibration: Fewer physical tests
Fast turnaround
Advantage of Finite Element tire models: Relative ease of calibration
o Coupon tests to determine cord material properties
o Continuous calibration not necessary
10
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
ValidationCo-simulation results validated against a standalone
Abaqus/Explicit simulation Vehicle travels over a bump 160 mm X 80 mm
Body and suspension components assumed rigid for simplicity and faster
turnaround
Results compared at the wheel centers as well as the reference nodes of
rigid bodies (control arm)
Subcycling ratio close to 300
11
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
ValidationInflation and Gravity loading using quasi-static Implicit
followed by Implict-Explicit co-simulation Import the tires into Abaqus/Explicit from the gravity loaded
configuration
Co-simulation between the models for pothole impact, fatigue
reference road, etc.
Gravity settling in Abaqus/Standard using quasi-static implicit
Remove TiresImport Tires
Implicit dynamics in Abaqus/Standard Explicit dynamics in Abaqus/Explicit
Co-simulation
12
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
ValidationLeft Wheel center
Comparison between Implicit-Explicit co-simulation and Standalone Explicit simulation at the left wheel centre.
(a) Longitudinal acceleration (b) Vertical acceleration (c) Vertical velocity (d) Vertical displacement
13
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
ValidationRight control arm center of gravity
Comparison between Implicit-Explicit co-simulation and Standalone Explicit simulation at the centre of gravity of the right
lower control arm. (a) Longitudinal acceleration (b) Vertical acceleration (c) Vertical velocity (d) Vertical displacement
14
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Use Cases: Curb Impact
• Stationary vehicle being impacted by a moving pendulum
— Quasi-static gravity settling and steering maneuvering performed in
Abaqus/Standard
— Assess damage to the suspension and steering system components
o Onset of plasticity expected in suspension components
— Body modeled as a substructure
— Event duration is very short (~50 ms)
15
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Use Cases: Pothole Impact
• Vehicle traveling at 30 km/h runs into a pothole— Onset of plasticity expected in parts of the Body
— Elastic-Plastic material used for the entire Body
— Event duration is moderately long (~500 ms)
16
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Use Cases:Fatigue Reference Roads
• Vehicle travels over roads paved with Belgian blocks— Event duration is long (~50s)
— Obtain road load data
— Body and suspension components modeled using substructures
17
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Template-based Suspension Modeling in Abaqus
18
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Template-based Suspension Modeling
Automotive Vehicle Suspension
Front Double Wishbone Type Rear Leaf Spring Type
19
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Suspension modeling in Abaqus
Unified CAE Analyses for Automotive Vehicle Suspension:
kinematics & compliance, vibration, and durability
http://www.ncac.gwu.edu/vml/models.html
20
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Abaqus/CAE Plug-in for suspension modeling
21
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Abaqus/CAE Plug-in for suspension modeling
Rigid Part -> Flexible Part
22
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Abaqus/CAE Plug-in for suspension modeling
Kinematics and Compliance Analysis
23
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Abaqus/CAE Plug-in for suspension modeling
Vibration Analysis
Rigid Model Partly Flexible Model
24
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Abaqus/CAE Plug-in for suspension modeling
Implicit Dynamics
Rigid Model Partly Flexible Model
25
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Case Study: Truck Suspension
Vehicle BODY
FRAME
FRONT SUS REAR SUS
Double Wishbone Leaf Spring
Body Mount
http://www.ncac.gwu.edu/vml/models.html
26
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Case Study: Truck Suspension
.
..
U-BOLT
Dummy Part
AXLE
REBOUND
CLIP (coupling)
Center Bolt(ignore in model)
1
2
n
..
.
In model,
*TIE is used assuming very
small relative movement.
Fish Plate
Pre-tension
27
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Case Study: Flexible Frame + Suspension forms
Vehicle Suspension Builder
28
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Case Study: Vibration Analysis
Vibration
~ 43 Hz
~ 1 Hz
29
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Case Study: Durability
Durability (Implicit Dynamics)
30
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
Summary
Co-simulation: A co-simulation technique combining the strengths of both implicit and explicit
solution techniques is implemented in Abaqus
o The methodology allows the implicit simulation to take time steps that are orders
of magnitude larger than the explicit time increments, without loss of accuracy
The results obtained using the co-simulation methodology for a full vehicle simulation
matches very well with those from a standalone explicit dynamic simulation
The schemes offers a powerful tool for full vehicle durability simulations
Rapid increase in compute power accessible to engineers is driving the shift towards
high fidelity system level simulation
Template-based suspension modeling Plug-in available for efficient building up certain suspension models
Leverages Abaqus functionality for easy to set up of K&C, Vibration and Durability
analyses
31
© D
assa
ult
Sys
tèm
es Ι
SG
L M
ichi
gan
RU
M,
Oct
ober
12,
2011
ABS example: Abaqus-Dymola co-simulation
LOGICAL AND PHYSICAL SIMULATION
Sophisticated hydraulics/state machine